Nov 17, 2022
Semiconductors: Is the Impactless Rebound Ending? (NVIDIA Earnings)
Ryunsu Sung
Micron throws cold water on the semiconductor rally
Micron (MU) had already reported earnings, but as the view spread in the market that memory semiconductors had bottomed, it released a press release to reset expectations. The company said it will further cut wafer starts by 20% in the fourth quarter.
Samsung Electronics and SK Hynix are also cutting wafer starts, but not to the same extent as Micron. As a result, Micron’s share in memory semiconductors is likely to shrink, while Samsung Electronics is expected to benefit.
Micron is also working toward additional Capex cuts. In calendar 2023, Micron now expects its year-on-year bit supply growth to be negative for DRAM, and in the single-digit percentage range for NAND.
For reference, Micron has already cut its 2023 capex (capital expenditures) by 50%. The press release implies there will be further capex reductions.
Micron now expects next year’s DRAM production volume to be negative versus this year.
This tells us that the current DRAM inventory problem is far more serious than many people think.
Over the past decade, bit growth in the memory market has exceeded 30% per year.
NVIDIA’s unjustified confidence
NVIDIA (NVDA) reported earnings.
- Revenue: $5.93B, -17% YoY, estimate: $5.79B
- Datacenter revenue – $3.83 billion, +30% YoY, consensus $3.79 billion
- Gaming revenue $1.57 billion, -51% YoY, consensus $1.32 billion
- Professional Visualization revenue $200 million, -65% YoY, consensus $349.1 million
- Automotive revenue $251 million, +86% YoY, consensus $242 million
- Adjusted gross margin 56.1% vs. 67% YoY, consensus 64.9%
- Adjusted EPS $0.58 vs. $1.17 YoY, consensus $0.70
On margins and expenses, the company issued fourth-quarter guidance that was better than consensus, but revenue guidance came in soft versus expectations.
- Revenue $6.00B +/- 2%, consensus $6.09B
- Gross margin 65.5% to 66.5%, consensus 65.3%
- Adjusted operating expenses $1.78 billion, consensus $1.83 billion
We expect our datacenter revenue to reflect early production shipments of the H100, offset by continued softness in China. In gaming, we expect to resume sequential growth with our revenue still below end demand as we continue to work through the channel inventory correction. And in automotive, we expect the continued ramp of our Oren design wins. All in, we expect modest sequential growth driven by automotive, gaming, and datacenter.
NVIDIA is saying it will grow in almost every segment, but given that overall revenue growth is expected to be modest, it appears that growth in datacenter revenue, which accounts for the largest share, will not be very strong.
Considering how datacenter has driven NVIDIA’s growth story for some time (and year-on-year growth is still high), a slowdown in datacenter revenue growth can only be seen as a negative for NVIDIA’s valuation multiple (P/E).
NVIDIA also recorded negative free cash flow for the first time in years.
Free cash flow -$156 million vs. $1.28 billion YoY, consensus +$2.63 billion
The fact that free cash flow turned negative despite revenue coming in above expectations is interpreted as a signal that inventories have piled up significantly. Yet NVIDIA claims that channel partners’ inventories are almost normalized.
We believe channel inventories are on track to approach normal levels as we exit Q4.
We booked an entry of $702 million for inventory reserves within the quarter. Most of that, primarily, all of it is related to our datacenter business, just due to the change in expected demand looking forward for China. So when we look at the datacenter products, a good portion of this was also the A100, which we wrote down.
Channel partners’ inventory levels may have normalized, but NVIDIA’s own inventory is surging. Inventory reached $4.45 billion, roughly double the $2.23 billion a year ago.
We are quickly adapting to the macro environment, correcting inventory levels, and paving the way for new products.
Jensen Huang
CEO Jensen Huang says he is adjusting inventory levels in line with the macro environment, but the numbers clearly tell a different story.
Now looking at our inventory that we have on hand and the inventory that has increased, a lot of that is just due to our upcoming architectures coming to market. Our Ada architecture, our Hopper architecture and even more in terms of our networking business. We have been building for those architectures to come to market and as such to say. We are always looking at our inventory levels at the end of each quarter for our expected demand going forward.
Given the TSMC capacity NVIDIA has committed to on the 5nm and 4nm nodes, its inventory is highly likely to keep rising.
In addition, NVIDIA took a $1.22 billion inventory charge in the previous quarter and a $700 million inventory charge this quarter. Combined, that is nearly $2 billion. If there had been no write-downs of inventory assets, NVIDIA’s inventory would have reached $6 billion, which is roughly three times the level a year ago.
Lastly, let’s look at the letter Meta CEO Mark Zuckerberg sent to employees as he carried out large-scale restructuring. Meta is one of the major customers that uses NVIDIA chips heavily for datacenters.
I’m currently in the middle of a thorough review of our infrastructure spending. As we build our AI infrastructure, we’re focused on becoming even more efficient with our capacity. Our infrastructure will continue to be an important advantage for Meta, and I believe we can achieve this while spending less.
Mark Zuckerberg
We are also conducting a thorough review of our infrastructure spending. As we build out our AI infrastructure, we plan to use AI more efficiently relative to scale.
Meta’s infrastructure will continue to provide a competitive edge, and I believe we can maintain that edge while spending less.
Three-line summary:
- Micron’s press release signals that the memory semiconductor inventory problem is more serious than investors perceive
- NVIDIA’s inventory continues to build, and without inventory write-downs it would be three times last year’s level
- There is a growing likelihood that major datacenter customers will cut back on their spending
Newsletter
Be the first to get news about original content, newsletters, and special events.