Last week I briefly mentioned the ARR growth pressure to Anthropic: https://t.co/WGWpRxTQKr. This was even before the astra release. The anxiety might be the reason behind the IPO delay.
This weekend, after talked to a few founders locally. I think astra is a much bigger threat to Anth. The cost per task is cheaper from the feedback and the service quality is much more stable during peak hours. People are voting with their feet
so I think it's urgent for anthropic to release a new model before the IPO, not only about the quality but also the inference efficiency.
That gap matters beyond API bills. If Astra can consistently reach similar quality with a much shorter reasoning chain, it potentially means better capacity utilization, lower cost per task, and less pressure during peak demand.
Fable is a great model. But if Anthropic needs ~2-3x the token generation to deliver comparable intelligence due to much longer thinking chain, the next release needs to reach to the next level of efficiency as much as raw capability to compete with OAI and retain the customers. Not to mention, OAI has secured way more compute power than Anth.
BofA's Simon Woo has put out a report, and there's some interesting content in it. To quote a portion:
- Recent checks with memory makers, spot market distributors, and OEMs point to firm 3Q ASPs: despite a rising LTA mix, 3Q ASPs for most DRAM products rose 20 to 30% quarter on quarter, and NAND rose more than 15%
- High single digit ASP increases expected in 4Q as well. Also a notable change: hyperscalers have signed new contracts agreeing to pay higher DRAM ASPs in 1Q27 than in 4Q26
- On 2H ASPs, the existing optimistic view is maintained, with no change to estimates. However, the stance on 1H27 has turned more positive given the persistent memory supply shortage and the aggressive race among OEMs and big tech to secure chips
- Accordingly, 2027 to 2028 DRAM/NAND ASP assumptions are raised by 8 to 12% and 2 to 3%, respectively. That said, ASPs are expected to decline around 10% in 2028 (DRAM down 5%, NAND down 13%), which is interpreted as a soft landing phase ahead of renewed growth in 2029 to 2030