I am of two minds on these stories. It is absolutely true that these are big cyclical commodity plays and pricing does peak. But it is also true that the commodity players have added very little new capacity, because they didn't believe either. They all thought the strength was re-stocking. But if you read the Intel (NASDAQ: INTC) (Cramer's Take) conference call transcript and you think they are being honest -- and I do -- you know we have just finished quarter one-and-a -half of actual build, not rebuild of inventories. And that is not enough to trigger massive capital expenditure on new equipment -- that is supposed to happen at the END of every cycle.

In other words, it is too soon and, alas, too cute. It is true that some of these stocks have run so much -- Western Digital -- that I don't want to own them. But it is also true that most haven't, when you look at longer-term charts.

How do we know it is early, besides Intel? How about the Linear Technology (NASDAQ: LLTC) (Cramer's Take) conference call, on which they said they are just beginning to consider new designs. That's the "buy Cadence (NASDAQ: CDNS) (Cramer's Take)" call, in case you are wondering.

How about the Xilinx (NASDAQ: XLNX) (Cramer's Take) interview I had last night, where the company was adamant about the strength of the mobile Internet tsunami?

I think selling these commodity plays is the wrong call. I think that buying a Cypress (NYSE: CY) (Cramer's Take) or an RF Micro (NASDAQ: RFMD) (Cramer's Take) or even a Micron or SanDisk is good, although I like the PLD biz of Xilinx more.

Most important, selling the components of PCs now, when the netbook is exploding in sales and China is ordering like mad, is too premature.

Hold on -- there is more coming.

Jim Cramer is co-founder and chairman of TheStreet.com. He contributes daily market commentary for TheStreet.com's sites and serves as an adviser to the company's CEO. At the time of publication, Cramer held no positions in the stocks mentioned.