2018 Predictions
Microsoft Google Amazon

Why Microsoft, Amazon Will Make Their Own AI Chips

By
Aaron Tilley
 |  Dec. 29, 2017 7:00 AM PST

Competing in the field of artificial intelligence requires a few key ingredients: huge amounts of data to train the systems, talented engineers and colossal amounts of computing power. That power is currently delivered, in large part, by graphics chips built by Nvidia, but there are signs that might change.

I predict that 2018 will be the year when every big tech company that is truly serious about its AI ambitions, including Microsoft, Facebook and Amazon, will attempt to build its own AI chips in an effort to gain an edge in computing power. The shift would echo steps already taken by Alphabet’s Google to build specialized chips, which have become crucial for servers in tech companies’ data centers that handle such AI tasks as image and speech recognition.

Get access to exclusive coverage
Read deeply reported stories from the largest newsroom in tech.
Latest Articles
 
Who Could Lead CNN if Zucker Leaves
Who Could Lead CNN if Zucker Leaves

As WarnerMedia CEO Jason Kilar overhauls the media firm to make it more digitally savvy and consumer focused, he is expected to zero in on its news and sports business, particularly CNN, for a shakeup, according to people familiar with the situation.

What could trigger a restructuring of the business is the possible departure of news and sports chairman Jeff Zucker. His contract isn’t up ...

Latest Briefs
 
Adobe Joining News Outlets to Combat ‘Deep Fakes'
Snap’s Revenue Surges and Sends Stock Soaring 20%
Netflix’s Subscriber Growth Slows Sharply
Stay in the know
Receive a summary of the day's top tech news—distilled into one email.
Access on the go
View stories on our mobile app and tune into our weekly podcast.
Join live video Q&A’s
Deep-dive into topics like startups and autonomous vehicles with our top reporters and other executives.
Enjoy a clutter-free experience
Read without any banner ads.
New Amazon Data: Retailer Wins Sales While Losing to Merchants on Listings
Data Point Policy E-commerce
New Amazon Data: Retailer Wins Sales While Losing to Merchants on Listings
Just 1% of the vacuum cleaners, immersion blenders and other household products listed on Amazon last year were sold by the retailer, with the rest coming from other merchants who use its platform.
New Data Show Reality Behind Self-Driving Truck Startup’s Funding
Exclusive Venture Capital Startups
New Data Show Reality Behind Self-Driving Truck Startup’s Funding
TuSimple, a prominent developer of self-driving semitrailer trucks that is worth more than $1 billion on paper, said last fall it had secured funding of $215 million from investors.
Censorship Drama Engulfs Facebook, Twitter: The Information’s Tech Briefing
The Briefing
Censorship Drama Engulfs Facebook, Twitter: The Information’s Tech Briefing
There’s no question that Facebook and Twitter face an extremely difficult task in trying to keep misinformation from polluting their networks, particularly now, in the weeks before the election.
Apple Unveils 5G Phones: The Information’s Tech Briefing
The Briefing
Apple Unveils 5G Phones: The Information’s Tech Briefing
Apple brought all its usual marketing hyperbole to today’s unveiling of its latest iPhones, with breathless exclamation for new colors, chip speed and the ceramic cover that Apple promises will be less breakable.
Cover Wall Street for The Information!
jobs@theinformation
Cover Wall Street for The Information!
If you’re interested in how the capital markets work, and you love the tech industry, The Information has a job for you.
The Rise of ‘Retention Rates,’ the Software Metric Investors Are Obsessing Over
Data Point Startups Markets
The Rise of ‘Retention Rates,’ the Software Metric Investors Are Obsessing Over
When database startup Snowflake filed for an initial public offering in late August, investors honed in on one measure of its performance: net revenue retention rate, which reflects a company’s ability to keep customers and get them to increase their spending over time.