Texas Instruments Made a Computer (& It Failed)

Founded nearly a hundred years ago, Texas  Instruments has undergone many transformations. Over the years they had smoothly  evolved from oil exploration company to military electronics supplier to  cutting-edge semiconductor maker. In the 1970s and 1980s, they sought to  make their most ambitious transformation yet - into a consumer electronics  company. And a computer-maker. It was a disaster. In this video,  we are going to take a look at TI’s ill-fated push into consumer  electronics, and th

e home computer. ## Beginnings Texas Instruments as we know it today emerged out  of a small oil and gas exploration service called Geophysical Service Inc or GSI. GSI was founded in 1930 by John Clarence Karcher and his good friend Eugene McDermott  to commercialize a new technology that Karcher invented for discovering  oil - reflection seismography. It kinda works like sonar for the  earth's insides. We use a dynamite explosion or an air gun to create seismic  waves. The waves bounce off thin

gs and we can read the responses to detect underground  structures like those that might contain oil. So the company's early business model  was to send crews of people out into the field on behalf of the big oil  companies to find new oil fields. ## GSI The technology worked, but 1930 marked the beginning of one of  America's worst economic depressions. Oil fell to historically low prices. GSI  responded by expanding internationally, sending crews to Venezuela, Ecuador, India, and Saudi Arabia.

This business did quite well  before Europe fell into war later in the decade. But back home, the exploration crews had  nothing to do. So management moved away from oil exploration/surveying and towards  production. Eventually the company found and developed several oil properties in  southwest Texas, Louisiana and Alabama. The oil companies did not enjoy the idea of their  former partner eventually competing against them. To defuse the tensions, Karcher and McDermott  proposed a restructuring

. Karcher would run the oil development business - named Coronado  - and GSI became Coronado’s subsidiary. ## Independence But the whole entity still lost money and clients still threatened to cancel  over conflicts of interest. So finally in 1941, Karcher decided to sell the  whole business including its 2,000 barrels of daily oil production for $5 million to Stanolind  Oil & Gas - forerunner of Amoco, which is now BP. To do right by GSI's customers, Stanolind  offered GSI's management the opti

on to purchase the subsidiary's assets  for $300,000. The potential buyers were Eugene McDermott and his partners Cecil  Green, H. Bates Peacock, and J. Erik Jonsson. It was not a slam-dunk decision.  The business was losing $10,000 a month. The war raging in Europe  had curtailed the international oil exploration business. Competition was  rising. The future seemed uncertain. Cecil Green was reluctant at first - even  writing a letter threatening to resign - but Jonsson eventually convinced him

to  change his mind. Green and Jonsson had to mortgage their personal properties  and stocks to pay their $75,000 share. ## War The buyout closed in November 1941. A month later was the attack on Pearl Harbor and the  United States entering World War II. The war meant all of GSI's crews  working abroad had to come home. One GSI crew doodlebugging on the  Indonesian island of Sumatra scrambled to escape the Port of Sunbaja ahead of  the approaching Imperial Japanese Army. With shots firing on th

e outskirts of  town, two members of the team boarded a small rubber transport ship which was then  hit by a Japanese torpedo, sinking all their possessions. They scrambled into a tiny lifeboat,  got picked up by a passing ship and made it home. With no choice, GSI pivoted to defense  technology. It came about a funny way. GSI cofounder McDermott receives  a call from a former colleague at Columbia University about building a thing  for the Navy. What’s that thing? Classified. That thing turned

out to be  a submarine detecting system called "Magnetic Anomaly Detector" or  MAD. It looks for submarines by finding disturbances in the earth's magnetic  field created by a large iron body. Despite first receiving a faulty circuit design  from the National Defense Research Committee, the machine they built worked  perfectly. Throughout the war, GSI became one of the nation's  biggest providers of MAD equipment. After the war ended in 1945, GSI was keen  to continue its pivot towards defense e

lectronics. Jonsson really liked one of the  military guys he worked with during the war, a 31-year old at the Navy's Bureau  of Aeronautics named Pat Haggerty. Haggerty wanted to try something  new so Jonsson offered him a job. ## L&M Haggerty joined as the new head of the Laboratory  and Manufacturing Division, or L&M Division. Established in the Spring of 1946,  the L&M Division numbered about 85 people. Their goal was to shore  up GSI's manufacturing chops to make them more competitive for 

military electronics business. Meanwhile the end of the war allowed GSI's oil  exploration crews to return to Saudi Arabia, Mexico and Brazil. These international revenues sustained the company as they built  out their manufacturing expansion. In 1947, GSI won their first large military  contract for a bomb sight. Two years later, they outbid Philco for their biggest  contract yet. To win the deal, GSI promised to deliver the APS-31  radar system - a thing they never built before - in just nine

months at a  fixed price. They nailed it, of course. ## Texas Instruments It was a time of great change for GSI. The Korean War was raging and the  company's L&M Division was rapidly growing from military contracts for air  radar systems and other equipment. The company was making millions in sales  from these products - not from oil. So in 1950, Jonsson and Haggerty felt  GSI needed a new name and organization that better reflected its current situation.  But what name? After a closed door sess

ion, they first came up with General Instruments,  which was announced in January 1951. But that name happens to be similar to  another defense contractor in New York called General Instrument. Singular.  People got confused and annoyed. So Jonsson threw a bunch of people into  a room and told them to come up with a new name. His only instructions  were that it should have "Texas" in the name. Not soon afterwards,  we got Texas Instruments, or TI. TI did an extensive restructuring, making  the G

SI oil exploration business one of its subsidiaries. That business ran  alongside all the semiconductor stuff for actually a pretty long time - it was  finally sold off to Haliburton in 1988. ## The Transistor In June 1948, Bell Labs announced  to the world the discovery of the first transistor - the point contact transistor. I covered this in a prior video. A transistor is  a device capable of either amplifying a signal, or controlling it like a switch.  The point contact transistor was a legit

imate scientific breakthrough, but not  rugged enough to be commercially successful. The later junction transistor - also known as a  bipolar transistor - offered better stability and reliability. It is made up of three alternating  sections of doped silicon or germanium. Those sections are called the emitter,  base, and collector. The boundaries between those areas create the "junctions"  that give the Junction transistor its name. The transistor will take in a small  voltage that is its input

signal, and use that to control a larger current flowing  through from the emitter to the collector. Advancements in growing pure germanium  and silicon crystal allowed Bell Labs to create a commercially viable junction transistor. Key to that effort was a Bell  Labs technician named Gordon Teal. He became one of the world's best at  growing crystals and selectively "doping" them with other elements like antimony to  create N and P-type silicon or germanium. ## Getting the Transistor Soon after

the 1948 Bell Labs announcement,  Haggerty quickly realized that the transistor would be special and decided to invest  in this semiconductor technology. Luckily for TI, there existed an avenue  to acquire it. AT&T's anti-trust deal with the US Government required them to license  out their technologies. So throughout 1951, TI tried to engage Bell Labs and acquire  a license to transistor production tech. The Bell Labs people did not hide their  skepticism about these random cowboys. Jonsson rec

alled being told by Bell executives  that TI was unlikely to succeed in transistors. They were not entirely wrong to be skeptical. Back  then, the value proposition for transistors was not so clear. Vacuum tubes were more battle-tested  and cost far less - like 60 cents each tube. Each Germanium transistor then had to be made  by hand - so they cost about $2-3 each. But TI persisted. They mailed a  $25,000 check to pay for the license fee and Haggerty even took graduate  physics courses at night

. In 1952, Bell Labs held a big symposium to discuss  how to produce transistors. TI attended. Upon returning, TI immediately set upon the work  of automating transistor manufacturing - assigning the task to one of their top guys  Mark Shephard. By the end of 1953, Shephard's team was turning out thousands  of grown junction transistors each month. ## Going to Silicon The beginning of 1953 saw another  big milestone - TI hired Gordon Teal. Haggerty and Teal had met at the 1952 Bell  Labs symposi

um. Teal was born in Dallas, Texas. After working for 22 years at Bell  Labs in New Jersey, he wanted to come home. TI offered him that way home along with a higher  salary and a chance to run his own research lab. So in January 1953, Teal became  a TI Assistant Vice President and founding head of TI Research Lab.  Soon as he started, he got to work building a team to produce very pure single  crystals of silicon for making transistors. Most of the early breakthroughs in solid state  physics occ

urred with germanium. But it soon became clear that germanium's lower melting point  and band gap made its transistors less reliable. When their internal temperatures hit  75 degrees Celsius, germanium-based transistors leak electrons and basically  fail. Silicon-based transistors on the other hand with their higher band gaps can  sustain temperatures of up to 130 degrees. ## Silicon Dreams The major challenge though would be  creating a silicon crystal pure enough. Silicon's higher temperature

and strong  reactivity mean that it easily absorbs impurities like oxygen as it is being  pulled out from the melt. Yet within a year and a half of Teal's arrival on April 1954,  TI produced silicon grown junction transistors. Teal recalls in a later oral history: > It took just the time necessary to  get somebody to help turn my suggestions into pieces of pulling equipment. I gave  suggestions ... so they built a crystal puller based on my suggestions and also  [from] looking at crystal pullers

that I was using in Bell Labs and had  already been using for sometime. So anti-climactic. Teal first announced the  breakthrough in May 1954, at a meeting of the IRE National Conference of Airborne  Electronics in Dayton, Ohio. He recalls: > During the morning sessions,  the speakers had unwittingly set the stage for us. One after the other  they had remarked how hopeless it was to expect the development of a silicon  transistor in less than several years. Teal listened, knowing that he had a

few  production silicon transistors literally in his pocket. Then he got up and gave his  presentation. Two-thirds of the way through, he dramatically revealed to  the audience what they had. Stunned, a person in the audience asked "Did you  say you have silicon transistors in production?" Teal said yes, and demonstrated their temperature  resilience by dumping a silicon transistorized amplifier into hot oil. It still worked, and  the audience members rushed to report the news. Texas Instruments

had actually not been  first to create a silicon transistor. Bell Labs' Morris Tanenbaum - Teal's colleague  - first fabricated them in January 1954. But Bell Labs failed to capitalize  on their achievement. They just sat on it. Either because they were already  heavily invested in germanium transistors, or it was not patentable, or just did  not think it was all that important. Meanwhile, TI rushed their  silicon transistors - TI 903, 904, and 905 - to market. And for several years  until 1958

, they were its only producers. Sales skyrocketed and margins were fantastic. The  silicon transistor became TI's big break. The US military was especially  interested in silicon transistor technology. The Soviets had recently sent  up Sputnik, and the silicon transistor’s ruggedness offered the American space  program a significant advantage. ## Pervasiveness As the silicon transistor neared its debut, Haggerty in the spring of 1954 decided  to make a transistorized radio. Making a consumer ele

ctronics device  might seem like a strange step for a company with a history rooted in transistors  and oil exploration equipment. And indeed, they just wanted to make transistors  and sell them to whoever wanted them. But as his protege Morris Chang would later say,  Haggerty always believed in the "pervasiveness" of semiconductors. These needed to be produced at  huge volumes. And to do that, TI needed a high volume sales channel. The military market was  great, but the commercial market was g

reater. Haggerty wanted to demonstrate to other  electronics companies still putting vacuum tubes into their products that transistor  technology had matured. And it was ready for them. ## The Radio Building what would eventually be  the Regency TR-1 was a regal ordeal. Haggerty initiated the program in Spring 1954,  and wanted the radio on consumer shelves by Christmas. The timeline was very tight. A  dozen engineers worked 10-12 hour days on a circuit board design, trying to balance the  right

number of transistors and their cost. For instance, the radio might play  beautifully with the newly introduced silicon transistors - but each of those  transistors cost $100. And even $30 for the whole radio was a bit much. So they stuck  with germanium. After this crash program, they had a 4-transistor board. But who was  going to manufacture and distribute it? Cofounder Cecil Green recalls: > Well, as I say, the major radio  companies weren't much interested, but we found that a small high f

idelity instrument  making company up in Indianapolis named Regency. At the time, Regency was only eight years  old. Founded by a few ex-RCA engineers, they were then making TV boosters. They took the  job because they were unafraid of anything new, and crafted a device big enough  for a good-quality dress shirt pocket and with a battery  life of about 20-30 hours. The radios went into production in November  1954. They sold at Neiman Marcus for about $49 each or $510 in 2021. It apparently  sol

d well enough to warrant successors. During its run, Regency's parent company chairman Ed Tudor said that half  of their sales came from radios. A few years later however, Japanese companies like  Sony, Hitachi, Toshiba, and Matsushita/Panasonic flooded the market - capturing 80% share  and driving the Americans out of the show. For TI, the financial impact was mixed. All in,  the program lost about $1.25 million - including assets bought and R&D spent from 1952  to 1955. But the device was a st

rategic success. Semiconductor sales rocketed, and in  1960 hit a staggering $200 million per year. One of those radio buyers was  a guy named Thomas Watson Jr., president of IBM, and the founder's  son. Impressed, he struck a highly coveted licensing deal with TI for their  transistor technology in December 1957. Then he wrote a memo telling his  sprawling organization, "After June 1, 1958, we will build no more machines with  electron tubes". Haggerty’s big bet paid off. ## Inventing the IC(?)

Over time, transistors gave  way to the integrated circuit. Scientists and industrialists have long envisioned  the commercial viability of mass-producing a miniaturized circuit with various transistors,  devices, and critically, their interconnects. In March 1959, TI announced that their  researcher Jack Kilby had produced the first working integrated circuit  or IC. Tested on September 1958, it had a transistor and a few other devices on a  slice of germanium. A silicon version came later. TI

called it the greatest thing since  the commercialization of the silicon transistor earlier in the decade.  And indeed, there was some value in demonstrating the feasibility of putting all  these devices on the same sliver of material. But the reality is that Kilby's device was  not very practical for commercial use. It didn't solve the IC's major issue of shrinking and  mass-producing interconnects between the devices. Six months after Kilby's invention,  Robert Noyce - future co-founder of In

tel but then working at Fairchild  - invents a fully integrated IC by building on top of his colleague Jean  Hoerni's breakthrough "planar process". Simply speaking, the planar process  leverages a special silicon trait - its ability to produce a layer of silicon  dioxide in the presence of water and heat. This thin silicon dioxide layer was  previously seen as a nuisance and removed. But silicon dioxide is an insulator,  and Hoerni recognized that they can use those insulating properties to bot

h  protect an IC's various components from contamination - as well as electrically  isolate it to prevent interference. Noyce then realizes that the silicon dioxide  layer can also prevent the metals in a metal wire connecting the IC's devices from  seeping into and poisoning the silicon. So he creates a step now known as  "aluminium metallization" - the back end process of laying down all of an IC's  connecting wires in a single wondrous step. The result is what we now call the  monolithic inte

grated circuit. It underperformed competing circuits of the  time, but it was far more reliable. And the fact that we can now print and produce  it in steps that do not correlate with its size or complexity allowed for rapid  miniaturization and cost scaling. So Kilby's thing came first, but Noyce's  thing was more practical. A long patent war thus naturally ensued, ending up  with the two guys sharing joint credit. And history largely concurs. Kilby and Noyce  would have won the Nobel together

in 2000 had Noyce not passed away in 1990 at the still-young  age of 62. I truly wish I could have met him. If you would like to deep dive  into this nerd rap battle, then there is a very very long  Wikipedia article you can read. ## The Calculators Throughout the 1960s, US Military spending on missile and rocket programs push  early IC investment. TI had over a hundred military contracts for products  like Transistor-Transistor Logic or TTL. But Haggerty still wanted more. Harkening back  to th

e Regency TR-1 radio, he decided to try the tactic again. Let us build something with these  ICs and sell it to the people to stimulate demand. In 1965 while boarding a business flight,  he suggests to Kilby the idea of a "personal computer" that can replace big heavy desktop  electric calculators and the slide rule. A slide rule predates even my time.  Wikipedia says that it is a mechanical calculator consisting of slidable  rulers for mathematical calculations. My father used to use one in  un

iversity and later handed it down to me when I was little. I  had no idea how to use it - still don't - but nevertheless treasured  it and it now sits in my study room. Anyway, Haggerty throws around a lot of dumb  ideas - one other suggestion on that 1965 flight was to make a lipstick sized dictation machine  and that was a stinker - but this one stuck. Kilby interpreted Haggerty's vague idea  to be a digital calculator product that would cost less than $100, can fit into a  coat pocket and do

addition, subtraction, division, multiplication, and maybe a square root. Kilby wrangles a team at TI's IC department  to build it, which is soon named the CAL-TECH. The working calculator was shown to Pat Haggerty  a year later in 1966. Made from solid aluminum, it was about four inches wide and  six inches long. It had 18 keys for input and printed out the results on a paper tape. TI shows the CAL-TECH to various companies and  Japan's Canon shows interest. The two partner up to produce the Ca

non Pocketronic  - released in April 1970 or 1971, sources differ. We largely consider it the  first commercial digital pocket calculator. Then in 1972, TI officially enters  the calculator market with the TI-2500 Datamath single-chip calculator,  the SR-10 scientific calculator, and others. The division became a $100 million  business, and a $200 million business the second year. TI previously had only wanted to  make transistors. But money changed that. ## Going Hard Into Consumer In 1973, Tex

as Instruments had  become a $1 billion company. Coming off the gleaming success of  the consumer calculator program, the company decided its future was to go deeper and  deeper into this consumer electronics strategy. The man chosen to spearhead this push was  Morris Chang. By 1967, he had risen to be general manager of TI's integrated circuits  division - its most important semiconductor division. He then became vice president  of all of TI's semiconductor businesses. After about six years or

so there, TI gave Morris  the consumer business portfolio. At a conference in 1974, Chang laid out the company's  thinking behind this push to consumer: > "The most important motivation for  vertical integration by semiconductor manufacturers is the opportunity for higher  growth resulting from increased value added for products where the principal function  is already performed by semiconductors" 30-35% of a calculator’s business  value is in its chip. Simply speaking, Morris said that by makin

g better semiconductors, you can bring more differentiated and  valuable products to the market at lower cost. Chang mentioned other products  with high market opportunity to be data terminals, electronic watches,  point-of-sale systems, and computers. The consumer market - then in the midst of  an Asian invasion - was a bloody battlefield, but TI had confidence. At a  shareholder conference in 1976, they officially announced their  entry into consumer electronics. ## Toys and Watches At the 197

5 Consumer Electronics Show in Chicago, TI displayed their next big  consumer product - digital watches. These watches had red LEDs or LCDs and  were powered by custom TI-made integrated circuits. In 1976, TI entered the market  with a bang - bringing out two high-end LCD lines selling for over $250 as well  as a low-end red LED watch for under $20. But the watch market is brutal.  Other semiconductor makers like National Semiconductor and Intel also took  the plunge into the watch business. The

year TI came out into the market, there were  over 40 different digital watch vendors. TI tried to undercut the market with a $9.95  watch, but margins were brutal. After few years, TI exited in 1981. Eventually  the traditional watchmakers reclaimed the market, focusing  on luxury and status-signaling. Overall, Chang did not look back fondly  on his years working on TI's consumer business. In retrospect, the calculator  success was perhaps a bit too much of a sweet success. New CEO Mark Shepar

d  refused to enter a business unless it had the potential to be a $200  million business within 3 years. Things that could have been  a decent niche business were set aside because the market wasn't  judged to be big enough. In the end, TI could only enter super-competitive, bloody  red markets - like the digital watch industry. And many of the businesses that TI did end  up entering required stretching outside TI's circle of competence. Throughout  the 1970s, TI made radio receivers, digital t

hermometers, home thermostats,  educational video games, and a CB radio. They even opened a retail outlet - the  TI Store - to showcase their products to the public. There were TI Stores  in Dallas, Italy, Berlin, and the UK. One product that Chang did seem to  be proud of was the "Speak & Spell", which used the world's first single-chip speech  synthesizer. The product grew out of efforts to find a commercial application  for higher capacity bubble memory. It was the hot toy of 1978, and even h

ad  prominent roles in the ET and Toy Story movies. ## The Computer TI's final, most costly foray - its 1812 invasion  of Russia - was into the home computer market. TI had been working on single-chip computers for a while. The Speak & Spell was built using  a TMS 1000 microcontroller - which also powered a variety of machines like microwave  ovens, calculators, video games, and so on. But TI largely disregarded the computer  and minicomputer market. Back then, such computers were at least the s

ize of cabinets, which limited their market potential.  How many such devices can you sell? What the people at TI and many other  companies did not foresee was the microcomputer - something we now call personal  computers. Kicked off by the Altair 8800 in 1975, the microcomputer industry spawned new names  like Apple, Commodore, and Tandy/Radio Shack. And when softwares like the  VisiCalc spreadsheet hit the market, microcomputers became more than just toys.  They became valuable record-keeping

and scheduling tools for small businesses  like accountants, dentists, and so on. TI realized that this market was  not going away. They knew what they had to do. So throughout 1979, the rumors  started to crop up everywhere. The giant, Texas Instruments, was going  into the home computer business. ## The TMS-9000 To power this computer, TI  needed a microprocessor. They wanted their offering to be the first 16-bit  home computer, meaning that its circuitry can handle 16 bits at the same time -

more memory  capacity allowing for more complex tasks. By comparison, the then-dominant  Apple II and Commodore PET ran 8-bit microprocessors. So TI's computer would  theoretically have a compute advantage. TI had long built custom minicomputers for  customers in the oil and scientific industries. TI VP Walden "Wally" Rhines later recalled in  an IEEE article that there was a "one company, one architecture" rule. So  they decided to adapt their minicomputers' 16-bit architecture  for this new mi

crocomputer chip. That turned out to be a mistake. The team  struggled to get this microprocessor - the TMS 9900 - out. Peripherals were limited  due to there being no 16-bit chips, necessitating the weird engineering  hack of adding an 8-bit port. But that compromised any benefits of the  processor having a 16-bit architecture. The chip also had 64 pins rather than  the traditional 40 pins, limiting its design wins outside of TI's own products. It would have done better had it been designed fro

m  the ground up as a microcomputer product. In late 1978, Rhines gave a presentation about  the TMS 9900 chip to a small group at IBM in Florida for an unnamed project - the IBM Personal  Computer. Unimpressed by the 9900's shortcomings, Boca Raton chose the 16-bit Intel 8088 for  their PC's microprocessor - a fateful choice. ## 800 Pound Gorilla Anyway, it didn't seem like a  huge loss at the time because Texas Instruments was the 800 pound  gorilla of the semiconductor industry. They were the

ones that Apple, Tandy,  and other guys feared would enter their space. Then it happened. In 1979, TI  brought out their 99/4 home computer. The computer's introductory price was  about $1,150 or $4,600 today - which was high for the time. Customers noted  the limited selection of peripherals and software - which focused on education,  personal finance and home management stuff. They also complained about  the calculator-style chiclet keyboard - reminds me of Apple's issues with their  own chic

let keyboard. Some things never change. At first, sales grew slowly. So the  TI team overhauled their strategy. They made a better keyboard, more peripherals, and cut the price by over half. The result  was the 99/4A Home Computer, or just the TI Home Computer. It sold to the retailer  for about $340 and to the consumer at $550. To sell the 99/4A, TI coopted many of the  microcomputer vendors' tactics. They cut deals with over 5,000 traditional  retail stores - Toys R Us, K-Mart, and such - not

the special computer  shops they used before. They hired thousands of teachers to educate  customers on how to use the machine. And they also orchestrated an ambitious  marketing campaign to get the word out. They bought a big TV ad campaign featuring  Bill Cosby - bad move looking back at it. ## Disadvantages But things soon spiraled out of control. A Texas Monthly article recalls how  TI originally wanted to do three computers - with a high and middle-end  computer alongside the low-end 99/4A.

However, manufacturing issues and personnel  squabbles derailed that effort. The other business groups felt the consumer group  was encroaching on their fiefdoms. The company's engineering-focused culture  also gave it a bad case of "Not Invented Here" syndrome. So they were unwilling to court  software partners and share profits with them, hurting their ability to build a strong third  party developer ecosystem around their computer. This especially hurt when IBM quickly captured  the professi

onal market with the PC. TI, Commodore, and Tandy thus found  themselves in the low-end consumer part of the pool - setting the  stage for a bloody price war. Nevertheless, TI invited  the battle. Years earlier, Texas had fought and won the calculator wars  - beating Commodore in that market so badly that they nearly filed for bankruptcy.  A story for another video, perhaps. Semiconductors have always been about dominating  the market with scale and low cost. Chips were the key part of the machi

ne, and TI is a chip  master. This should be their game right? Looking back at it, entering into a price war  was the wrong thing to do. They could have gone for an education niche like with the  calculators. The 99/4A was a high-quality, well-made computer. Better designed, more capable, and more costly to make than its competitors.  TI should never have gotten into that octagon. ## War But they did and this time  Commodore made them pay. Their low-end Commodore Vic-20  computer's simpler desig

n combined with their production lines in  Asia allowed for far deeper cuts than TI could sustain. Commodore's  executives knew this. TI's did not. The Vic-20 hit the market at $300. So TI matched. They cut retail prices on the 99/4A down  from $525 at that time to $449 to $300. Already at this point, the computer is selling for  very little margin. Then Commodore cuts the Vic-20 price down to $200. That pricing situation  persisted throughout the spring of 1982. In September 1982, the 99/4A is

selling at  $300, but the Vic-20 is about $100 cheaper. That is when TI added rebates to cut the  price down to $200 - a shot across the bow. Sales surged. During the second half of 1982, TI was selling 99/4As like hotcakes - 150,000  units a month in 12,000 retail outlets. TI's PC business had blown up into a $200 million  business almost overnight. Books and software were being written for the computer.  Large user support groups sprouted up. But even as sales skyrocketed, TI refused to  open

up the architecture to and share profits with third party software developers like Apple  or Commodore did - building a lockout chip into its computer just like Nintendo did. TI  really did need those software profits, but it was another missed opportunity  to build something long-term. In an attempt to keep up, in December 1982  Atari cut their 8-bit home computer the Atari 400 to under $200. They were a distant  third in the segment and fading fast. But Commodore had more to give. In January 1

983  - just as TI were feeling themselves at the peak of the home computer market - Commodore cut their  top-end Commodore 64 computer down to just $400; and their low-end Vic-20 to under $130. TI in turn matched by cutting retail  prices on the 99/4A to just $150. The retailers screamed foul. Cosby joked about how easy it was to sell a  computer when TI paid you $100 to buy one. Atari could not sustain this kind  of competition. In February 1983, they abruptly fired a quarter of their US  staff

and moved all of their factories to Asia. Then TI blinked and stumbled. As  Atari melted down, TI discovered a serious issue with the 99/4A's power  supply transformer - not referring to the LLM architecture - that could  potentially cause users to get shocked. Nothing yet happened, but nevertheless TI reported  the issue and decided to shut down the line to rework the product. For a few critical months in  1983, the 99/4A was not available on the shelves. In April 1983, Commodore delivered the

  final blow. They cut the Vic-20 down to $99, which was about what the 99/4A cost TI  to make. TI had no response to this, and for the first time they did not match. Commodore also slashed the  Commodore 64's retail price as well as prices on software - the  only thing sustaining TI's margins. Commodore had won. Assembly lines kept  cranking out TI 99/4As, but now the retailers started returning unsold inventory.  Sales projections missed again and again. TI's accountants realized that the home

computer  division and its price wars had lost the company a staggering $200+ million - causing a total  $119 million loss in the second quarter of 1983. Such a loss, even a giant like TI could not  sustain and the stock shed a third of its value. The home computer division head Bill  Turner resigned. And in October 1983, TI announced their exit from the home  computer business. Taking into account both operating losses and write-offs,  they had lost a staggering $680 million. Looking back agai

n at it, pulling out so abruptly from the home computer market might  not have been the right thing to do too. It was a good computer. There was a  lot of goodwill built up by enthusiasts. With the new 99/2 in development - internally  nicknamed the Apple Killer - a pivot into education to challenge Apple Computer might  have sustainable. But alas it was not to be. In the wake of this painful withdrawal, TI made  a final grasp to the professional market with the TI Professional Computer. But IBM

was already  there with their PC, and TI - again with the Not Invented Here syndrome - refused to make their  product fully IBM-compatible. It failed too. ## Conclusion TI's failed foray into the consumer  electronics field had lasting implications. It torpedoed Morris Chang's previously  fast-rising career at TI. While he was promoted to Senior Vice President, he  lost his shot at the coveted CEO spot. Put out to pasture, he left for  General Instrument and then after that - Taiwan. Perhaps a

bit sour over  TI's ventures into consumer electronics, he later proposed a semiconductor  company that would never do such a thing. The home computer failure not only lost  TI their place in a growing market, but also snuffed out the only big end  user demand market they had for their own microprocessors. Their digital  logic business was never the same. TI has since largely retreated back  to their core competencies of making semiconductors - with a focus on  analog, mixed signal and digital s

ignal processing ICs. They remain one of  America's biggest semiconductor companies. Though they still make calculators. The TI 99/4A remains a classic,  and one of the most famous home computers in American history.  Its fans remember it quite fondly. For more information about  Texas Instruments' history, I highly recommend this 2005 book by Caleb  Pirtle III. They made it for the company's 75th anniversary so it's a bit biased  but the pictures and stories are great.