Snapshot · The a16z Show
Gavin Baker predicts that within the next 3 years, a number of high-profile custom AI chip (ASIC) programs will be canceled, especially if Google begins selling TPUs externally.
Where this was said
At 24:44 · chapter starts 22:35
Baker delivers his most detailed technical analysis in the chip segment, mapping a competitive landscape that most investors misread. NVIDIA's real enemy isn't AMD — it's Google's TPU, the only credible alternative for AI training and possibly the best inference alternative today. Broadcom is playing a fascinating enabling role, offering hyperscalers like Meta an Ethernet-based fabric that theoretically competes with NVIDIA's NVLink stack, alongside custom ASIC development services [1] — Gavin Baker "NVIDIA's real competitor isn't AMD — it's Google's TPU. Broadcom is playing a clever game, offering hyperscalers custom Ethernet-based fabr…" 23:00 . The catch: custom silicon is brutally hard. Google took three generations to get the TPU right. Baker predicts that within three years, most high-profile custom ASIC programs will be canceled, especially if Google begins selling TPUs externally — a development widely rumored on social media. He notes that Amazon's Annapurna silicon team is the most talented at any hyperscaler, and Trainium 3 will be significantly better than Trainium 2, but Google ultimately controls the TPU and can pull it away from Broadcom at any time. AMD will always exist as the market's necessary second source.
GPT-5 is a smaller, more economical model — not a frontier capability push. Using it to claim scaling laws are dead is completely wrong. Baker's message: don't conflate efficiency optimization with a capability ceiling.
NVIDIA started as a semiconductor company, became a software company through CUDA, then a systems company through rack-level solutions, and is now architecting at the data center level. Baker says Jensen Huang is one of the two best CEOs he has ever encountered — and he's playing a very strong hand.
NVIDIA's real competitor isn't AMD — it's Google's TPU. Broadcom is playing a clever game, offering hyperscalers custom Ethernet-based fabrics and ASIC alternatives to NVIDIA's NVLink stack. But Baker predicts most custom ASIC programs will be canceled within 3 years, especially if Google starts selling TPUs externally.
Google required three chip generations to make the TPU viable — a cautionary benchmark for other companies attempting custom silicon.
Bhanu and his team built approximately 50 free tools to attract search traffic, each linked back to SiteGPT.
With AI coding tools like Cursor, Bhanu can now create a new free marketing tool in less than 5 minutes by referencing existing tools.
Bhanu filters Ahrefs keyword results to show only those with a keyword difficulty below 10, making them realistic ranking targets for any decent website.
Bhanu sets a minimum search volume of 1,000 monthly searches when selecting keywords to target with free tools.
PropGPT averaged 20 downloads per day right after launching on the App Store through influencer marketing.
Eyal and Yali shut down all marketing and spent 4 months completely rebuilding PropGPT from scratch.
PropGPT has accumulated over 40,000 total downloads since launch.
PropGPT's large language model (AI) operating costs are just $20 per month, and the cost is continually falling.
Ad-based monetization works well for game apps where users spend extended time in-session, as seen with Grid and Wordle.
We use essential and analytics cookies to run Vuci. To understand how the site is used: Privacy Policy.
Install Vuci on your phone
Add it to your home screen for a faster, app-like experience.
Install Vuci on your phone
Tap the Share button, then “Add to Home Screen”.
A new version is available
Reload to get the latest Vuci.