logo
Why a16z's ‘speedrun' accelerator is playing a new game

Why a16z's ‘speedrun' accelerator is playing a new game

Yahoo02-04-2025
The term "speedrun" is usually reserved for Super Mario.
"In games, a speedrun is a runthrough where you try to beat the game as fast as you possibly can," said Josh Lu, principal at Andreessen Horowitz. "A canonical example is Super Mario Brothers, where there's a set number of levels and you have to beat them all as fast as you can. You don't get extra credit for getting more coins, larger Marios, flower Marios. You just get credit for how quickly you beat the game."
And, in startups, certain principles of this can apply.
"Speed really matters in early-stage startups, the speed by which teams execute, the speed by which they answer the important questions," said Lu. "Can you get to product-market fit? Do you know your audience? Are you able to hire great people to fill out your team? Speed matters."
Lu helps run a16z speedrun, an accelerator the firm started in 2023 in San Francisco and Los Angeles to focus on games. Since its founding, speedrun has backed more than 120 companies and, this week, the accelerator opened applications for its 5th LA cohort, bumping up investment per company to $1 million from $750,000. For the last cohort, more than 6,000 companies applied, but only around 40 were accepted, marking an acceptance rate of slightly less than 1%.
I'd been seeing a16z speedrun all over social media over the last year, so when they offered Fortune an exclusive look at the accelerator at a key moment, we said yes—this March, speedrun expanded its mission to focus more broadly on "the intersection of tech and sort of entertainment," said Lu.
"We think games are at the heart of a lot of entertainment, and what makes games special are the same things at the heart of entertainment experiences," he told Fortune. "These are all things that cause people to feel, to play with each other, to evoke emotion. We consistently saw that the best game-makers were just really great storytellers."
So far, a16z has deployed more than $100 million in capital towards speedrun—to be sure, that's small by a16z standards: Per its latest SEC filing, a16z has about $74.74 billion in assets under management. Still, $100 million over a short time has been used to back a striking variety of companies, including online child safety startup k-ID, content creation platform Hedra, Echo Chess maker Echo Chunk, and AI consumer company Dream Voyage. The accelerator also has previously backed various blockchain-based companies, like Voldex and Mythical Games, and now's a time when crypto is very "in" in Washington, as is a16z itself.
"When we select a company, that's our belief in the ability of the team to build something really great," said Lu. "If we tried to over-rotate towards where we think policy is going, we have less visibility on that and it's not part of our process."
Coco Chen, cofounder and CEO at Dream Voyage, was part of the second speedrun cohort, and said she wanted to be around other people specifically building in consumer.
"I haven't done any other accelerator so it's hard to say, but as I've talked to friends who've been through others, I think the programming seems similar—other programs have, say, a weekly speaker session," said Chen. "But I do think the network is different. When I hang out with my friends who work in B2B SaaS they care about different things, like long sale cycles and finding key decision makers. When people build games for consumers, they ask different questions."
Sami Ramly, founder and CEO at Echo Chess, was in the third speedrun cohort, and says that he still uses key things he learned, in a way that actually all ties back to the accelerator's name.
"I had this preconception that when you speedrun through something, there are fundamental prerequisites, like facing the exact same challenge 1,000 times before you finally speed through it," said Ramly. "That's speedruns in the game world. And I think it's a great metaphor, but doesn't fully do the program justice. What I've learned is that speedrunning isn't just about improvement, practice, or repetition, that you can learn to separate good insights from a pattern of success, without having to go through them 1,000 times."
See you tomorrow,
Allie GarfinkleX: @agarfinksEmail: alexandra.garfinkle@fortune.comSubmit a deal for the Term Sheet newsletter here.
Nina Ajemian curated the deals section of today's newsletter. Subscribe here.
This story was originally featured on Fortune.com
Orange background

Try Our AI Features

Explore what Daily8 AI can do for you:

Comments

No comments yet...

Related Articles

Exclusive-Nvidia working on new AI chip for China that outperforms the H20, sources say
Exclusive-Nvidia working on new AI chip for China that outperforms the H20, sources say

Yahoo

time12 minutes ago

  • Yahoo

Exclusive-Nvidia working on new AI chip for China that outperforms the H20, sources say

By Liam Mo and Fanny Potkin BEIJING/SINGAPORE (Reuters) -Nvidia is developing a new AI chip for China based on its latest Blackwell architecture that will be more powerful than the H20 model it is currently allowed to sell there, two people briefed on the matter said. U.S. President Donald Trump last week opened the door to the possibility of more advanced Nvidia chips being sold in China. But the sources noted U.S. regulatory approval is far from guaranteed amid deep-seated fears in Washington about giving China too much access to U.S. artificial intelligence technology. The new chip, tentatively known as the B30A, will use a single-die design that is likely to deliver half the raw computing power of the more sophisticated dual-die configuration in Nvidia's flagship B300 accelerator card, the sources said. A single-die design is when all the main parts of an integrated circuit are made on one continuous piece of silicon rather than split across multiple dies. The new chip would have high-bandwidth memory and Nvidia's NVLink technology for fast data transmission between processors, features that are also in the H20 - a chip based on the company's older Hopper architecture. The chip's specifications are not completely finalised but Nvidia hopes to deliver samples to Chinese clients for testing as early as next month, said the sources who were not authorised to speak to media and declined to be identified. Nvidia said in a statement: "We evaluate a variety of products for our roadmap, so that we can be prepared to compete to the extent that governments allow." "Everything we offer is with the full approval of the applicable authorities and designed solely for beneficial commercial use," it said. The U.S. Department of Commerce did not respond to a Reuters request for comment. FLASHPOINT The extent to which China, which generated 13% of Nvidia's revenue in the past financial year, can have access to cutting-edge AI chips is one of the biggest flashpoints in U.S.-Sino trade tensions. Nvidia only received permission in July to recommence sales of the H20. It was developed specifically for China after export restrictions were put in place in 2023, but company was abruptly ordered to stop sales in April. Trump said last week he might allow Nvidia to sell a scaled-down version of its next-generation chip in China after announcing an unprecedented deal that will see Nvidia and rival AMD give the U.S. government 15% of revenue from sales of some advanced chips in China. A new Nvidia chip for China might have "30% to 50% off", he suggested in an apparent reference to the chip's computing power, adding that the H20 was "obsolete". U.S. legislators, both Democratic and Republican, have worried that access to even scaled-down versions of flagship AI chips will impede U.S. efforts to maintain its lead in artificial intelligence. But Nvidia and others argue that it is important to retain Chinese interest in its chips - which work with Nvidia's software tools - so that developers do not completely switch over to offerings from rivals like Huawei. Huawei has made great strides in chip development, with its latest models said to be on par with Nvidia in some aspects like computing power, though analysts say it lags in key areas such as software ecosystem support and memory bandwidth capabilities. Complicating Nvidia's efforts to retain market share in China, Chinese state media have also in recent weeks alleged that the U.S firm's chips could pose security risks, and authorities have cautioned Chinese tech firms about purchasing the H20. Nvidia says its chips carry no backdoor risks. Nvidia is also preparing to start delivering a separate new China-specific chip based on its Blackwell architecture and designed primarily for AI inference tasks, according to two other people familiar with those plans. Reuters reported in May that this chip, currently dubbed the RTX6000D, will sell for less than the H20, reflecting weaker specifications and simpler manufacturing requirements. The chip is designed to fall under thresholds set by the U.S. government. It uses conventional GDDR memory and features memory bandwidth of 1,398 gigabytes per second, just below the 1.4 terabyte threshold established by restrictions introduced in April that led to the initial H20 ban. Nvidia is set to deliver small batches of RTX6000D to Chinese clients in September, said one of the people.

CNBC Daily Open: OpenAI CEO, who sparked AI frenzy, worries about AI bubble
CNBC Daily Open: OpenAI CEO, who sparked AI frenzy, worries about AI bubble

CNBC

time14 minutes ago

  • CNBC

CNBC Daily Open: OpenAI CEO, who sparked AI frenzy, worries about AI bubble

There's a bubble forming in the artificial intelligence industry, according to OpenAI CEO Sam Altman. "Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes," Altman said. "I'm sure someone's gonna write some sensational headline about that. I wish you wouldn't, but that's fine," he added. (Apologies to Altman.) Altman's AI company is currently in talks to sell about $6 billion in stock that would value OpenAI at around $500 billion, CNBC confirmed Friday. In another conversation, Altman warned that the U.S. may be underestimating the progress that China is making in AI. Given the above premises, should investors be more cautious about OpenAI? Altman was not posed this question, but one wonders whether his opinion would also be "yes." Outside pure-play AI companies, the money is, likewise, still flowing. Intel is receiving a $2 billion injection of cash from Japan's SoftBank. It's a much-needed boost to the beleaguered U.S. chipmaker. Intel has fallen behind foreign rivals such as TSMC and Samsung in manufacturing semiconductors that serve as the brains for AI models. But going by Altman's views, the investment in Intel might not be a good bet by SoftBank CEO Masayoshi Son. Not everyone agrees with Altman, of course. Wedbush's Dan Ives told CNBC on Monday that there might be "some froth" in parts of the market, but "the actual impact over the medium and long term is actually being underestimated." And Ray Wang, research director for semiconductors, supply chain and emerging technology at Futurum Group, pointed out that the AI industry is not heterogeneous. There are market leaders, and then there are companies that are still developing. In the real world, bubbles delight because they reflect their surroundings in a play of light. But the bubble Altman described could be one doesn't show the face of its meeting paves the way for trilateral talks with Putin. At the White House meeting, the U.S. president also discussed security guarantees for Ukraine — which would reportedly involve a purchase around $90 billion of American weapons by Kyiv. Intel is getting a $2 billion investment from SoftBank. Both companies announced the development Monday, in which SoftBank will pay $23 per share for Intel's common stock. Shares of Intel jumped more than 5% in extended trading. The artificial intelligence market is in a bubble, says Sam Altman. Separately, the OpenAI CEO said he's "worried about China," and that the U.S. may be underestimating the latter's progress in artificial intelligence. U.S. stocks close mostly flat on Monday. The three major indexes made moves that were less than 0.1 percentage pointsin either direction as investors await key U.S. retail earnings. Asia-Pacific markets were mixed Tuesday. SoftBank shares fell as much as 5.7%. [PRO] Opportunities in one area of the European market. Investors have been pivoting away from the U.S. as multiple European indexes outperform those on Wall Street. But one pocket of Europe still remains overlooked, according to analysts. American money pours into Europe's soccer giants as club valuations soar European soccer is a bigger business than ever, with clubs in the continent's five top leagues raking in 20.4 billion euros ($23.7 billion) in revenue in the 2023-2024 season. American investors have been eyeing a piece of that pie. U.S. investors now own, fully or in part, the majority of soccer teams in England's Premier League. That now includes four of the traditional Big Six clubs, with Chelsea, Liverpool, Manchester United and Arsenal all attracting U.S. investment.

OpenAI Rolls Out $5 ChatGPT Go Plan Exclusively for India
OpenAI Rolls Out $5 ChatGPT Go Plan Exclusively for India

Bloomberg

time44 minutes ago

  • Bloomberg

OpenAI Rolls Out $5 ChatGPT Go Plan Exclusively for India

OpenAI is launching an affordable new subscription plan in India, seeking to expand its artificial intelligence offerings in one of the world's biggest internet service markets. The new ChatGPT Go plan costs 399 rupees ($4.60) per month and allows customers to generate more images and interact with the chatbot more frequently than the free version, which will continue to be available. Among the other upgrades are higher limits on file uploads, longer memory for more personalized responses and more in-depth analytics.

DOWNLOAD THE APP

Get Started Now: Download the App

Ready to dive into a world of global content with local flavor? Download Daily8 app today from your preferred app store and start exploring.
app-storeplay-store